Going serverless

Streaming Data with AWS Kinesis and Lambda

Maksim Pecherskiy

Data Engineer

Last chapter

Lambda

Serverless

• Servers managed by cloud provider
• Pay per execution
• Automatic scaling
• Lower memory and exec time limits
• Great for quick, targeted functions
• Responding to an event with a callback

Servers (Traditional)

• Servers managed by user
• Pay per hour of running the machine
• Manual scaling
• Higher memory and exec time limits
• Great for long running code executions
• Training a machine learning model

1. Trigger

2. Handler

3. Layer

4. Destination

Powerful combination

Data Transformation

API

Alerting

Alexa

Sample handler

import json

def lambda_handler(event, context):

    return {
        'statusCode': 200,
        'body': json.dumps('Hello from Lambda!')
    }

Review

Review

Data Transformation

API

Alerting

Alexa

Review

Serverless

• Responding to an event

Traditional

• Training an ML model

Review

Let's practice!

Streaming Data with AWS Kinesis and Lambda